Verdict machinery: On the need to automatically make sense of test results